Yes, a trend arrow might work really well. I like that idea.
Another thought is that the V/S indication has a logarithmic presentation, (or expanded centre scale), but the ILS is linear.
If we made the ILS markers logarithmic/expanded centre scale, it would be easier to notice deviation trends. So half the marker travel from centre could be, say 0.25° LOC deviation. The next 1/4 of scale would go up to say 1° LOC dev, the last quarter up to 2.5° dev.
But I think your idea is better.
